Laid-open specification DE 10 2010 063 984 A1 describes a sensor system, comprising a plurality of sensor elements and a signal processing device, wherein the signal processing device is configured in such a way that the output signals from the sensor elements are evaluated jointly.
The invention is based on the object of proposing a sensor system which enables a relatively high level of accuracy with respect to the evaluation and/or processing of the output signals and/or data from sensor elements.
The sensor elements may be designed in such a way that they detect at least partially different primary measured variables and use at least partially different measurement principles.
Expediently, the sensor system is arranged in a vehicle, in particular a motor vehicle, particularly preferably an automobile.
The fusion filter may be designed in such a way that function steps are implemented successively, each function step being assigned time information, wherein, in the function steps, in each case data/values with respect to physical variables are processed and, at defined times, fusion data sets are provided which each provide a relative value of the physical variable and information on the data quality in respect of each physical variable which is processed and/or calculated by the fusion filter.
the fusion filter, at least with respect to one of the physical variables, has, internally, an absolute value and/or internal value of this physical variable, wherein the absolute values and/or internal values of the preceding function steps with respect to one or more physical variables are stored and kept ready for a defined time span.
It is expedient for, depending on the time information of the input values, a physical variable, of the fusion filter, these input values to be computed in the fusion filter by virtue of being compared and/or computed with the absolute value and/or internal value of the same physical variable, the time information of which, i.e. the time information of the absolute value and/or the internal value, substantially corresponds to the time information of the input value.
The absolute value and/or internal value of the same physical variable with which the input value is compared and/or computed from the two temporally next, with regard to the time information, absolute values and/or internal values which are stored and kept ready, with this in particular being the absolute value and/or internal value to be interpolated with time information which is slightly more recent and the absolute value and/or internal value with time information which is slightly older than the time information of the assigned input value.
It is expedient for time information to be assigned to each of the input values of the fusion filter, as a result of which these input values of physical variables each have a priority, wherein, as part of the pending function step, the input values of physical variables of the fusion filter are computed with internal values of the fusion filter of the same physical variables, which correspond in terms of priority and/or with regard to the time information substantially to the respective time information and/or the priority of the respective input variable.
The signal processing device is designed in such a way that the priority of the presently pending function step is defined by the time information of the input value of a physical variable with the most recent time information, wherein the input values of physical variables whose time information is older are extrapolated onto the priority of the most recent time information and then computed with internal values of the fusion filter.
The signal processing device to be designed in such a way that the defined time span in which data/information/values are stored and kept ready is defined by the maximum delay of the input data of the fusion filter, which are determined from the maximum difference between the time information of the measurement of a physical variable and the time information of a presently pending function step of the fusion filter, wherein the maximum delay is in particular defined as being greater than or equal to the maximum expected delay of a measurement of a satellite navigation system which is part of the sensor system.
It is expedient for the signal processing device to be designed in such a way that the defined time span in which data/information/values are stored and kept ready is defined by a defined time span which is different for some or each physical variable.
The sensor system comprises a satellite navigation system, which provides a synchronization pulse and/or a synchronization method, with which the time information of the fusion filter is synchronized with the time information of the satellite navigation system.
It is expedient for the satellite navigation system to provide distance data in each case between the assigned satellite and the vehicle or a variable which is dependent thereon and velocity information data in each case between the assigned satellite and the vehicle or a variable dependent thereon, which distance data and velocity information data are transmitted as input values to the fusion filter, wherein these input values of physical variables of the satellite navigation system are computed, in time-corrected fashion, with internal values of the fusion filter by virtue of, depending on the time information of the input values, the internal values of these physical variables being extrapolated if the time information of the input values of these physical variables are older than time information of the presently pending function step of the fusion filter, wherein, for the extrapolation, in particular a gradient of the last two values of such a physical variable is used if these two values are based on two successive measurements between which no measurement of the same measured variable has failed.
The fusion filter may be designed in such a way that it takes into consideration at least one model assumption, which is based on the fact that the offset values and/or change values and/or correction values and/or error values between one or more function steps of the fusion filter change only to a limited extent or negligibly within the defined time span for which, in particular, absolute values and/or internal values of physical variables are stored.
The invention is based on the concept that, in order to compute the redundant measurement data of various sensors or sensor elements or optionally additionally a satellite navigation system which are measured at different times with one another in a fusion filter and therefore to be able to make use of the advantages of redundancy in sensor fusion, it is advantageous to compute measured variables and/or internal values of the fusion filter which are present, belonging to the same measurement time, in order to keep the error between outdated measurement data which form the input values of the fusion filter and present measurement data namely the values with respect to the measurement data which are present at that time in a pending function step of the fusion filter, as small as possible.
The fusion filter is designed in such a way that the comparison values or internal values of physical variables of the fusion filter are buffer-stored over a certain time or a defined time span, which should in particular cover all of the delay times of all of the input data, and then the appropriate internal value from the past is selected for the comparison for the data fusion and is computed with an input value. Specifically, this can be limited, for example, by the sampling/sample rate/measurement times, and particularly preferably a certain interval remains between internal values of the filter and measurement data. Alternatively, the values are finely interpolated between two sampling operations of the filter to such an extent that a valid measured value appropriate for the time of the measurement data is present. If the internal values from fusion and measurement data or input values, in each case of the same physical variable, are now present at the same time or with substantially the same assigned time information, they can be computed/fusioned with one another easily. Expediently, a comparison of the data from the fusion with those of the measurement data is performed, and from this the corrected values which update the values of the fusion filter are then determined.
The signal processing device may be designed in such a way that all of the input values of the fusion filter are each stored together with the assigned time information for a defined time span and that the absolute values and/or internal values of the preceding function steps of the fusion filter in respect of one or more physical variables are stored together with the respectively assigned time information for a defined time span. It is expedient that the priority of the presently pending function step is defined by the time information of the input value of a physical variable with the most recent time information. Expediently, depending on the time information of the input values, a physical variable, of the fusion filter, these input values are computed in the fusion filter by virtue of them being compared and/or computed with the absolute value and/or internal value of the same physical variable, the time information thereof, i.e. the time information of the absolute value and/or the internal value, substantially corresponding to the time information of the input value. The result of this computation has the priority or relates to the time of the time information which was assigned to the input variable and the absolute value or internal value. On the basis of this result, one or more function steps are implemented by the fusion filter, with which in each case input data and absolute values or internal values which are more recent by one function step are in each case calculated forward in time until the event of the function step is present which has the priority or to which time information is assigned which corresponds to the priority of the function step presently pending at the beginning of this calculation process, which has been defined by the time information of the input value of a physical variable with the most recent time information.
A priority is understood to mean absolute or relative time information or alternatively a position in a temporal sequence, alternatively in particular a combination of the two.
The fusion filter may be in the form of a Kalman filter, alternatively a particle filter or alternatively an information filter or alternatively in the form of an “unscented” Kalman filter.
The fusion filter may be designed in such a way that the fusion data set comprises, as value of the at least one physical variable, a relative value, in particular an offset value and/or change value and/or correction value and/or error value.
It is expedient for the relative values of the respective physical variables of the fusion data set to be correction values, to each of which scattering information or scattering or scattering degree, in particular a variance, is assigned as information relating to the data quality of said correction values.
The fusion filter may be designed in such a way that the value of at least one physical variable of the fusion data set is calculated on a direct or indirect basis from sensor signals from a plurality of sensor elements, wherein these sensor elements detect this at least one physical variable in a direct or indirect manner, with redundancy. This redundant detection is particularly implemented as direct or parallel redundancy and/or as analytical redundancy, from computationally derived or deduced variables/values and/or model assumptions.
The fusion filter may be in the form of a Kalman filter which iteratively implements at least prediction steps and correction steps and at least partially provides the fusion data set. In particular, the fusion filter is in the form of an error state space extended sequential Kalman filter, i.e. in the form of a Kalman filter which particularly comprises linearization and in which error state information is calculated and/or estimated and/or which operates sequentially and in the process uses/takes into consideration the input data available in the respective function step of the sequence.
It is expedient for the sensor system to have an inertial sensor arrangement, comprising at least one acceleration sensor element and at least one rotation rate sensor element, and for the sensor system to comprise a strapdown algorithm unit, in which a strapdown algorithm is implemented, with which at least the sensor signals of the inertial sensor arrangement are processed in particular corrected navigation data and/or driving dynamics data, on the basis of the vehicle in which the sensor system is arranged.
It is particularly preferable for the strapdown algorithm unit to provide its calculated navigation data and/or driving dynamics data to the fusion filter directly or indirectly.
The sensor system may have an inertial sensor arrangement, which is designed in such a way that it can detect at least the acceleration along a second defined axis, in particular the transverse axis of the vehicle, and at least the rotation rate about a third defined axis, in particular the vertical axis of the vehicle, wherein the first and third defined axes form a generating system, and in the process are in particular oriented perpendicular to one another, wherein the sensor system also has at least one wheel rotation speed sensor element, in particular at least or precisely four wheel rotation speed sensor elements, which detect the wheel rotation speed of a wheel or the wheel rotation speeds of in each case one of the wheels of the vehicle and in particular additionally detect the direction of rotation of the assigned wheel of the vehicle in which the sensor system is arranged, wherein the sensor system additionally comprises at least one steering angle sensor element, which detects the steering angle of the vehicle, and wherein the sensor system furthermore comprises a satellite navigation system, which is designed in particular in such a way that it detects and/or provides the distance data in each case between the assigned satellite and the vehicle or a variable dependent thereon and velocity information data in each case between the assigned satellite and the vehicle or a variable dependent thereon.
Particularly preferably, the inertial sensor arrangement is designed in such a way that it can detect at least the accelerations along a first, a second and a third defined axis and at least the rotation rates about these first, second and third defined axes, wherein these first, second and third defined axes form a generating system, and in the process are in particular in each case oriented perpendicular to one another.
It is preferable for the inertial sensor arrangement to provide its sensor signals to the strapdown algorithm unit and for the strapdown algorithm unit to be designed in such a way that it at least calculates and/or provides at least corrected accelerations along the first, the second and the third defined axes, at least corrected rotation rates about these three defined axes, at least a velocity with respect to these three defined axes, and at least one position variable, as measured variables and/or navigation data and/or driving dynamics data from the sensor signals of the inertial sensor arrangement and in particular at least fault state information and/or variance and/or information on the data quality which is assigned to a sensor signal or a physical variable and is provided by the fusion filter.
It is expedient for the sensor system to be designed in such a way that in each case at least one sensor signal and/or a physical variable, as direct or derived variable of the inertial sensor arrangement and/or the strapdown algorithm unit, of the wheel rotation speed sensor elements and the steering angle sensor element, in particular indirectly via a vehicle model unit, and of the satellite navigation system, in this case in particular distance data in each case between the assigned satellite and the vehicle or a variable dependent thereon and velocity information data in each case between the assigned satellite and the vehicle or a variable dependent thereon, are provided to the fusion filter and taken into consideration by the fusion filter in the calculations it performs.
It is particularly preferable for the vehicle model unit to be designed in such a way that the speed along the first defined axis, the speed along the second defined axis and the rotation rate about the third defined axis are calculated from the sensor signals of the rotation speed sensor elements and the steering angle sensor element.
It is very particularly preferable for the vehicle model unit to be designed in such a way that, for the calculation, a least-squared-error method is used for solving an overdetermined system of equations.
It is expedient for the vehicle model unit to be designed in such a way that, in its calculation, it takes into consideration at least the following physical variables and/or parameters    a) the steering angle of each wheel, in particular detected by the steering angle sensor for the two front wheels, wherein the model assumption whereby the steering angle of the rear wheels is equal to zero or the steering angle of the rear wheels is additionally detected is used,    b) the wheel rotation speed or a variable dependent thereon for each wheel,    c) the rotation direction of each wheel,    d) the dynamic radius and/or wheel diameter of each wheel, and    e) the track width of each axle of the vehicle and/or the wheelbase between the axles of the vehicle.
The signal processing device is preferably designed in such a way that the fusion filter calculates and/or provides and/or outputs the fusion data set at defined times.
The fusion filter is preferably designed in such a way that it calculates and/or provides and/or outputs the fusion data set independently of the sampling rates and/or sensor signal output times of the sensor elements, in particular the wheel rotation speed sensor elements and the steering angle sensor element, and independently of temporal signal or measured variable or information output times of the satellite navigation system.
It is expedient for the signal processing device to be designed in such a way that, over the course of a function step of the fusion filter, the newest information and/or signals and/or data available to the fusion filter of the sensor elements, in particular of the wheel rotation speed sensor elements and the steering angle sensor element, are always updated, in particular asynchronously, directly or indirectly, in particular by means of the vehicle model unit and the satellite navigation system directly or indirectly, sequentially and/or are recorded by the fusion filter and taken into consideration in the calculation of the assigned function step of the fusion filter.
It is preferable for the sensor system to have a standstill identification unit, which is designed in such a way that it can identify a standstill of the vehicle and, in the event of an identified standstill of the vehicle, provides information from a standstill model at least to the fusion filter, in this case in particular the information that the rotation rates about all of the three axes have the value zero and at least one position change variable likewise has the value zero and in particular the velocities along all three axes have the value zero.
It is preferable for the signal processing device to calculate and/or use a first group of data of physical variables, whose values relate to a vehicle coordinate system, and wherein the signal processing device additionally calculates and/or uses a second group of data of physical variables, whose values relate to a world coordinate system, wherein this world coordinate system is suitable in particular at least for describing the orientation and/or dynamic variables of the vehicle in the world, wherein the sensor system has an orientation model unit, with which the orientation angle between the vehicle coordinate system and the world coordinate system is calculated, wherein the orientation angle between the vehicle coordinate system and the world coordinate system is calculated in the orientation model unit at least on the basis of the following variables: the velocity with respect to the vehicle coordinate system, the velocity with respect to the world coordinate system and in particular the steering angles.
It is expedient for the following terms to be used synonymously, i.e. have the same meaning when implemented technically: offset value, change value, correction value and error value.
Error state information is preferably understood to mean error information and/or error correction information and/or scattering information and/or variance information and/or accuracy information.
The term variance is preferably understood to mean scatter, wherein in particular in the case of a general fusion filter, said filter in each case assigns scatter or a scatter value to each value of a physical variable of the fusion filter, and in the case of a Kalman filter as the fusion filter, in each case a variance is assigned to each value of a physical variable of the fusion filter.
It is expedient for the first, second and third defined axes on the basis of a coordinate system of the vehicle in which the sensor system is implemented to be defined as follows: the first defined axis corresponds to the longitudinal axis of the vehicle, the second defined axis corresponds to the transverse axis of the vehicle, and the third defined axis corresponds to the vertical axis of the vehicle. These three axes in particular form a Cartesian coordinate system.
It is preferable for the fusion filter to be designed in such a way that its data, in particular the physical variables or the data of the physical variables of the fusion data set, are divided into blocks which always have a constant size and which are processed iteratively in any desired order in the fusion filter, i.e. the fusion filter implements a sequential update with respect to its input data. In this case, the fusion filter is particularly preferably designed in such a way that the filter equations are matched, with the result that the computational result of the sequential update in each step of the fusion filter is an update, i.e. a data update, for all measured variables of the input data of the fusion filter.
The sensor system is expediently arranged in a vehicle, in particular in a motor vehicle, particularly preferably in an automobile.
The sensor system is preferably designed in such a way that data of the satellite navigation system, in particular position data, are assigned timestamp information, which substantially describes the measurement time of these data. The timestamp information of the respective datum of the satellite navigation system is provided jointly with this respective datum to the fusion filter and taken into consideration in the internal calculation in the fusion filter.
Preferably, in addition, such timestamp information is likewise assigned to the data of further or all of the sensor elements and/or the inertial sensor arrangement, which timestamp information is likewise provided with the respective datum to the fusion filter and is taken into consideration in the internal calculation in the fusion filter. Expediently, the respective timestamp information is generated by the satellite navigation system itself with respect to the data of the satellite navigation system.
It is preferable for the respective timestamp information to be generated by the signal processing device in the case of the additional timestamp information of the further sensor elements and/or the inertial sensor arrangement, in particular depending on the time measurement of the satellite navigation system.
Preferably, a function step of the fusion filter comprises at least one prediction step and a correction step. The fusion filter is in this case formed iteratively and performs iteratively, one after the other, function steps. In particular, data or values or signals are input within each function step of the fusion filter, i.e. input data are taken into consideration, i.e. data or values or signals are also output, i.e. provided as output data.
The fusion filter is preferably designed in such a way that the fusion filter implements a plurality of update steps within a function step, wherein these update steps relate to loading or use or updating of input data or signals. The fusion filter runs in particular sequentially through all of the input variables or input signals and checks in each case whether new information/data are present. If this is the case, this information or data is transferred into the filter or the information/data are updated in the filter, and if this is not the case the present value is maintained and the filter checks the next input or the next input variable or the next input signal.
The strapdown algorithm unit preferably provides at least absolute values of physical variables, in particular absolute values for the acceleration, the rotation rate, the velocity, in this case in each case in relation to the three axes, to the vehicle and/or world coordinate system, and a position and the orientation angle. The values with respect to these variables are in this case particularly preferably all provided by the strapdown algorithm unit as corrected values/variables.
It is expedient for the inertial sensor arrangement to clock and/or trigger the fusion filter, in particular each fusion step which is implemented by the fusion filter is triggered by the inertial sensor arrangement or at least one output signal or output datum.
It is preferable for the strapdown algorithm unit to be designed in such a way that it has a start vector of physical variables and/or a start value of the position, in particular with respect to the start of the sensor system, particularly preferably after each time the sensor system is switched on. The strapdown algorithm unit particularly preferably receives this start vector and/or this start position via the fusion filter from the satellite navigation system.
It is expedient for the data of the fusion filter, in particular the fusion data set thereof, to image a virtual sensor or correspond to such a virtual sensor.
The term sensor elements is preferably understood to mean the wheel rotation speed sensor elements, the at least one steering angle sensor element, the sensor elements of the inertial sensor arrangement and in particular additionally also the satellite navigation system.
If, in general, a variable and/or value is specified in respect of the three defined axes, it is preferable for this to be intended with respect to the vehicle coordinate system and/or the world coordinate system.
It is expedient for the fusion data set, which comprises values of the physical variables, to comprise a relative value, for example a correction value, also referred to as offset value, and in particular to be provided to the strapdown algorithm unit. In accordance with the example, this respective correction value results in each case from the accumulated error values or change values which are provided by the fusion filter.
In addition, the invention relates to the use of the sensor system in vehicles, in particular motor vehicles, particularly preferably in automobiles.